specification-architect

The Specification Architect skill is functionally legitimate: it prescribes a structured, traceability-first process and enforces evidence-backed research. There is no direct sign of intentionally malicious code in the provided fragment. However, operational and supply-chain risks are present: the mandatory, high-volume external fetching and the unspecified local 'validation' script and 'execute' action are escalation points that could enable arbitrary code execution or data exfiltration if misused. Recommended mitigations before use: (1) require human review and signature before any local script execution or 'execute' action, (2) implement a trusted-source allowlist and provenance verification for WebFetch targets, (3) adopt explicit secrets handling policies (use vaults, never log credentials), and (4) include the validator script in the reviewable repo and/or perform a static audit prior to execution. With those safeguards, the skill can be used safely to produce high-assurance documentation.